1 of 29

Creating Serendipity�in Science

Personal anecdotes → general lessons

Creating the tinder for that spark of ideas

Mark V. Albert

Director, Biomedical AI Lab

Chair, Information Science

University of North Texas

1

2 of 29

“Creating Serendipity”?

Serendipity:

How did you have that idea?

How did the team come together?

How did you get that opportunity?

A lot of accomplishments for academics or industry professions can seem like they came together by random chance, but there is an active approach to increase the chance of these “random occurrences”

2

3 of 29

Serendipity, defined, is quite apt

Christian Busch views serendipity as "active luck", where chance encounters and human action come together. A missed flight or a casual walk in the park can lead to new friendships, interests, or even career opportunities.[5][6]

While serendipity in popular usage is often understood as a matter of pure chance, scientific discussions emphasize the crucial role of human agency—recognizing, interpreting, and acting upon unexpected opportunities. This interaction between chance and conscious action has been a key theme in areas such as creativity, leadership, innovation, and entrepreneurship.[6][7][8]

3

4 of 29

Series of anecdotes → general principles

This starts as a collection of personal stories

… focused on the scale of accomplishment, then tracing back to the first steps

… there’s bragging: otherwise where is the evidence that this is a good path and you should listen?

My examples are related to Health Informatics

… my research is Health-related but my administrative hat represents my college and department

… but the principles apply to many domains

Your goal with these stories, or any other such stories you hear over the years, is to extract a few common themes that can apply to you, or those around you. We will bring them together explicitly at the end

4

5 of 29

Stories

  1. Motion Analysis Center research with Shriners Children’s Hospital System
  2. Year-long AI consulting gig with Deloitte
  3. Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  4. Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  5. Gesture-to-speech recognition system
  6. Instinctiv.com startup during graduate school

5

6 of 29

First example, Shriners Hospital collaboration

Over $600,000 worth of grants to UNT through the Shriners Children’s Hospital system

Shriners Children’s Hospitals System grant Jan 2026 - Dec 2027

“Predicting and Optimizing Long-Term Outcomes in Pediatric-Onset SCI: A Machine Learning Approach” ($114,363 to UNT) awarded

Shriners Hospitals for Children system grant Jan 2023 - Dec 2025

“Improving Orthopaedic Outcomes through Machine Learning” ($289,770 to UNT)

$400k total - recently, an additional ~$100k in a 4th year funded extension to the 3 year grant

Shriners Hospitals for Children system grant Jan 2020 - Dec 2021

“Machine learning prediction using quantitative gait analysis data to suggest and support interventions for pediatric patients with movement disorders” ($111,218 for UNT)

6

7 of 29

Overall outcomes prediction approach

7

8 of 29

8

9 of 29

9

At the last Shriners Children’s Hospital system State of Science meeting there were

  • 14 Shriners Hospitals
  • Georgia Tech & Emory
  • And UNT!

10 of 29

Shriners Children’s Hospital - Chicago

Where did it all begin?

  • Random guy just stopped by asking for early analysis with an excel file! (Adam Graf, Senior Motion Analysis Lab Engineer at Shriners Chicago)
    • Looped in Joe Krzak and Karen Kruger over time as plans developed
  • Third time's a charm:
    • An unsuccessful cycle of undergraduates on a first effort
    • An unsuccessful small development grant
    • A later resubmission that was funded! (with Joe Krzak as PI)
  • Visits, in person, to Chicago Shriners to work on these iterations
    • Working on-site at key times to know the team, their needs, and be clear on commitment
  • Monthly and Biweekly zoom meetings ever since!

10

11 of 29

Stories

  • Motion Analysis Center research with Shriners Children’s Hospital System
  • Year-long AI consulting gig with Deloitte
  • Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  • Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  • Gesture-to-speech recognition system
  • Instinctiv.com startup during graduate school

11

12 of 29

AI consulting gig with Deloitte

Consulting on the Cognitive Analytics in at Deloitte in Chicago

Met Director of the Cognitive Analytics team, Plamen Petrov, when I invited him to a Loyola University Panel about career trajectories - room with only 25 students.

Brief conversations there led to an awareness how we could help each other.

A few lunches afterward, mostly talking shop. Taught a 2-day workshop together, started writing a book together…

12

13 of 29

AI consulting gig with Deloitte, more valuable than pay

Opportunities snowballed after that

I met Brian Ray through his presence on the Deloitte team, at the time ChiPy organizer and working under Plamen

  • Led a symposium for 150 python programming professionals sharing research from the lab

Great result: Connected Brian Ray to Mark Shapiro, a former colleague* who was transitioning from academia to industry.

Shapiro has advised many graduates since then (including my first PhD student who ended up working at the same company)

13

* Albert MV, Toledo S, Shapiro M, Kording KP (2012) Using mobile phones for activity recognition in Parkinson’s patients. Frontiers in Neurology 3:158

14 of 29

Stories

  • Motion Analysis Center research with Shriners Children’s Hospital System
  • Year-long AI consulting gig with Deloitte
  • Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  • Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  • Gesture-to-speech recognition system
  • Instinctiv.com startup during graduate school

14

15 of 29

Rehabilitation Technologies and Outcomes Research

Working with Arun Jayaraman started as side project with primary advisor

Over years (2010 - 2013) transitioned 100% into his lab

1-2 paper published with his team each year ever since (shows up 20 places in my CV)

Shaped my entire career trajectory - started by just trying to be helpful!

15

16 of 29

Stories

  • Motion Analysis Center research with Shriners Children’s Hospital System
  • Year-long AI consulting gig with Deloitte
  • Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  • Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  • Gesture-to-speech recognition system
  • Instinctiv.com startup during graduate school

16

17 of 29

Toddler Activity Recognition

Ongoing collaboration with Northwestern Feinberg School of Medicine

Did wearable activity recognition for patients →Soyang Kwon at Lurie Children’s Hospital interested in tracking toddlers to understand obesity. Both of us early career.

Now Toddler tracking!

3 MS students at Loyola worked on the project in natural cycles

Few publications, but stuck with Soyang throughout by seeing potential and need!

Now consulting on a successfully awarded NIH R01 to Soyang based on those efforts, funding a PhD student for this coming year to continue the work

17

18 of 29

Stories

  • Motion Analysis Center research with Shriners Children’s Hospital System
  • Year-long AI consulting gig with Deloitte
  • Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  • Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  • Gesture-to-speech recognition system
  • Instinctiv.com startup during graduate school

18

19 of 29

Results of a visit by Hannah Thompson

19

20 of 29

Flexible gesture recognition

to enhance communication

for motor impaired individuals

20

21 of 29

Results of application driven research - a step closer!

Grant: C-STAR Pilot Project (role: subcontract) Jan 2024 - May 2025

Center for Smart Use of Technologies to Assess Real-world Outcomes (C-STAR) award for “Pilot development and assessment of a precision gesture-to-speech system for speech-impaired individuals with limited mobility” ($37,125 to UNT through competitively awarded subcontract from NIH grant #P2CHD101899)

Papers:

  1. Rafiq RB*, Shi W, Albert MV (submitted) KARL: Knowledge-Attentive Representation Learning for Wearable Hand Gesture Recognition in Motor-Impaired Individuals. AISI 2026
  2. Rafiq RB*, Shi W, Albert MV (2024) Wearable Sensor-Based Few-Shot Continual Learning on Hand Gestures for Motor-Impaired Individuals via Latent Embedding Exploitation. IJCAI 2024
  3. Rafiq RB*, Karim SA*, Albert MV (2023) An LSTM-based Gesture-to-Speech Recognition System. 2023 IEEE 11th International Conference on Healthcare Informatics (ICHI). 430-438. doi: 10.1109/ICHI57859.2023.00062.

21

22 of 29

Stories

  • Motion Analysis Center research with Shriners Children’s Hospital System
  • Year-long AI consulting gig with Deloitte
  • Rehabilitation Technologies and Outcomes at the Shirley Ryan AbilityLab
  • Toddler activity recognition with Lurie Children’s Hospital, Now Northwestern
  • Gesture-to-speech recognition system
  • Instinctiv.com startup during graduate school

22

23 of 29

Instinctiv.com

2008-2009: Worked 10 hrs/week with two others on Instinctiv.com (the CEO and CTO) and 5 Bulgarian software engineers. Traveled to Sofia, Bulgaria over Thanksgiving that year to meet rest of team.

Started by answering questions my officemate posed during graduate school. Then “I should pay you for this” and he did.

2009: left to do a postdoc

2012: acquired by soundcloud for $10M

(I didn’t see any of that $$$)

CTO leveraged the situation quite well. �Founded Hyperscience

23

24 of 29

What can we infer from these examples?

24

25 of 29

What can we infer…

BE COLLEGIAL!!!

Open to help, and be helped, by others.

25

26 of 29

What can we infer…

  • Aim to be in positions with “doers” around you and gravitate toward them.
  • Status is less important than you may think, in the long run.
  • Character, attitude, and focus are more important than you may realize.

26

27 of 29

What can we infer…

  • Have a calling/expertise that people can associate you with.
  • Not just something interesting to you, but useful to others.

27

28 of 29

What can we infer…

  • Connections do matter, follow up with people.
  • Stick around as they grow.
  • Okay to focus on practical goals of connection and be professional.

28

29 of 29

In summary, some guiding principles

BE COLLEGIAL!!! Open to help, and be helped, by others.

Aim to be in positions with “doers” around you and gravitate toward them. �Status is less important than you may think, in the long run.�Character, attitude, and focus are more important than you think.

Have a calling/expertise that people can associate you with.�Not just something interesting to you, but useful to others.

Connections do matter, follow up with people. Stick around as they grow.�Okay to focus on practical goals of connection and be professional.�

29